Another thing missing is the procedure when multiple members recuse themselves from a decision. It seems important:
(a) that decisions are always made by more than one individual
(b) that applicants are not disqualified from receiving grants on the grounds of (e.g.) having COIs with multiple fund managers.
It is also worth thinking about COIs when it comes to board composition. Having multiple board members that are more isolated romantically/socially/geographically from the Bay Area scene would make the fund more robust to COIs, particularly if the standards for recusal are raised.
A fund member had some substantial past romantic and/or sexual relationship with a potential grantee that ended more than a year ago, or is a metamour of an applicant in an ongoing relationship
...
This is not sufficient cause for recusal, but should be made public when the fund member decides to not recuse themselves and the grant is made
I'm not sure information about past romantic/sexual relationships/metamourships should ever be made public.
An external board should be available to deal with situations like this - they can then either insist on recusal, or judge the COI not to be a problem.
I agree this would be a very bad outcome, and I'm surprised you think it is a possibility. Individuals/organisations should not be precluded from receiving LTF grants due to their personal lives, particularly if LTF continues to be the major funder for certain types of grants. I think the onus is on LTF to find a way of managing COIs that avoids this, while also having a suitably stringent COI policy.